12
\$\begingroup\$

I am trying to find a short on my board. So far, I have used my fluke to measure the resistance to gain some clues but everywhere I measure 4 ohm, so that didn't work.

I am thinking of trying to push <2V up to couple hundred mA with a power supply (current limited) make it work for a minute or so to find the heat location but not so sure that is safe. Is this the only way to do this?

\$\endgroup\$
  • 3
    \$\begingroup\$ Is this a bare PCB or with components loaded? \$\endgroup\$ – tehnyit Jul 6 '11 at 9:52
  • \$\begingroup\$ I was wondering the same as tehnyit. The symptoms (4\$\Omega\$ everywhere) seem weird for both. \$\endgroup\$ – stevenvh Jul 6 '11 at 11:04
  • 2
    \$\begingroup\$ Sounds like his meter is off by 4 ohms or so. \$\endgroup\$ – Mike DeSimone Jul 6 '11 at 12:15
  • 1
    \$\begingroup\$ @Mike - do you mean it's actually 8 \$\Omega\$? :-) \$\endgroup\$ – stevenvh Jul 6 '11 at 12:25
  • 2
    \$\begingroup\$ No, I mean that if he touched the meter's probes together, the meter would read something close to 4 ohms. \$\endgroup\$ – Mike DeSimone Jul 6 '11 at 12:27
13
\$\begingroup\$

A surer method is to apply current and then measure voltage drop across sections of track with a sensitive voltmeter. Many multimeter (even cheap ones) have a 200 mV (milliVolt) range which, with a 4 digit display gives 100 uV (microVolt) resolution (not accuracy, but that's another issue). If you pass say 100 mA along a track then you need 1 milliOhm of track to drop 100 uV.

  • \$(R = \frac{V}{I} = \frac{0.0001}{0.1} = 0.001)\$

If you place two probes at either end of a track section the polarity of the voltage drop will show you the direction of current flow. You can literally follow the current around the board. You can find the point at which current leaves a track - voltage drop per track length will be smaller or of opposite polarity past a point of current departure.

If that's not sensitive enough, many meters have a 200 uA range with a resolution of 0.1 uA (!). Due to issues of meter internal resistance and specific implementation one or other of these ranges may work best for you, but either way you have a useful short-circuit-current tracer.

Note that the voltage range is less liable to be damaged when doing this. When using a current range, if you probe two locations which are not on the same track and which have a significant voltage difference it is easy for far more than 200 uA to flow and, depending on meter circuitry, damage may occur.

You can set the test current to a useful minimum value by using a current limited supply to apply a current to a section of track and then observing the results with your meter. Current may be limited to a value that works well enough with you test equipment.

Whether a given test current is "safe" depends on your circuitry and is not able to be stated for a general situation. If you use a voltage and current limited supply and set the supply voltage to just high enough to produce a short circuit current suitable for testing then you will know the maximum energy that is able to be dissipated anywhere. For example, if you set Imax to 100 mA and Vsupply to 2 Volts then total available energy = V x I = 2 x 0.1 = 200 mW (milliWatt). That level of dissipation MAY cause problems in specific devices but is not liable to cause thermal damage to almost anything.

\$\endgroup\$
  • \$\begingroup\$ This is great.. I have learned something valuable here.. I will use this method.. \$\endgroup\$ – Frank Jul 7 '11 at 2:33
  • \$\begingroup\$ I've used this method with pretty good success on tracking down components that were damaged and drawing too much current. You do need a good multimeter to use this technique. A handheld DMM probably won't cut it unless you apply a large current. \$\endgroup\$ – CHendrix Nov 21 '16 at 16:49
  • \$\begingroup\$ I have also just used this method with great success to locate an invisible short between 2 MCU pins. I used a max current of 100mA and was able to trace the drop in voltage around the board using a cheapo handheld DMM until I found the short. Thanks very much, saved hundreds of pounds by not having to buy a milliOhm meter or IR camera. \$\endgroup\$ – donturner Feb 23 '19 at 23:28
7
\$\begingroup\$

True unintended shorts on a PCB are quite rare these days. I think it's been 10 years since the last time I saw stray copper on a board that wasn't supposed to be there. The first thing you should do is check your design files with the assumption that you messed up, not the board house. Did you remember to run the electrical rules check? The design rules check? That's what they are called in Eagle. Your package may have different names, but the concepts are the same. In the 10 years I haven't seen a short due to PC house, I have made a few mistakes like accidentally had two nets the same name after some editing of the schematic and the like.

If it's really the board house's fault, then apply some current and watch the voltage drops as Russell has said. It sounds a lot better in theory than in practise. Often the voltages you will see won't seem to make sense.

My first job out of school was for HP in New Jersey. Half the division made power supplies. One of the supplies was for electroplating. It put out 5V at 100s of amps, and was the size of a small washing machine that rolled on casters. My tech was fond of finding shorts by connecting the shorted nets to one of these supplies. There was a loud pop, some smoke, but the scortch marks and hole the size of a dime would give you a pretty good idea where the short was.

\$\endgroup\$
  • \$\begingroup\$ I can say I have definitely had this happen, even from first-class US PCB Fabs (Advanced Circuits, in this case). I wound up tracking it down to a tiny, tiny hairline copper whisker between two close pads. I used a thermal camera, and applied power and looked for localized heating. This is probably not a widely available approach, though. \$\endgroup\$ – Connor Wolf Aug 31 '13 at 14:26
3
\$\begingroup\$

I have successfully tracked down board-shorts using a thermal camera.

You apply power, and look for a temperature rise. It works quite well.
You do need an infrared camera (or a infrared thermometer, and lots of patience), though.

\$\endgroup\$
0
\$\begingroup\$

Very often there is a problem in finding the fault location in the board, if the closure is in the voltage supply circuits. Of course you can remove parts of the series with a circuit board, but it takes a lot of time, especially if the board is complicated. My idea is, that would be to search for the fault location with the help of Hall sensor. In the short circuit must be increased current density, and consequently a large magnetic field.I have an idea how to make such a device. It's easy and cheap

\$\endgroup\$
  • \$\begingroup\$ So you're saying a hall sensor that is small enough to approximately be the size of the PCB traces, thus able to distinguish between current flowing on one trace as opposed to another? Also sensitive enough to sense the current through a trace on a presumably low-power PCB? Interesting. \$\endgroup\$ – Anindo Ghosh Aug 31 '13 at 15:03
  • \$\begingroup\$ How does an idea you haven't tried yet help to answer the question? \$\endgroup\$ – Dave Tweed Aug 31 '13 at 15:49
  • \$\begingroup\$ This seems like it could be an interesting answer if you completed the thought... \$\endgroup\$ – Chris Stratton Aug 31 '13 at 16:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.